Tag: T shirts

Speculators rapidly adjust portfolios in response to shifting interest rate expectations and escalating global conflicts. The leading cryptocurrency recently experienced a notable decline while major United States stock indices suffered significant losses. This dual downturn highlights a broader risk-off strategy among individuals anticipating tighter monetary policy and higher energy costs. These factors create a highly volatile environment demanding careful analysis of underlying metrics rather than superficial valuation movements. We must evaluate specific numbers driving these asset classes to understand true directional momentum.

The top digital token currently trades at US$78,510.84, down 0.77 per cent over the last 24 hours. This instrument slightly underperforms a relatively flat broader financial landscape. Such divergence indicates crypto-specific macro positioning drives current valuation action rather than general equity trends. We see a weak correlation between traditional safe havens and stocks during this specific window.

The S&P 500 moved down just 0.06 per cent while Gold gained 0.29 per cent over the same period. Participants currently treat the primary decentralised network strictly as a rate-sensitive risk instrument. They reduce exposure to non-yielding speculative tokens as risk-free government bond yields climb. This behaviour confirms that digital currency ecosystems operate with unique internal dynamics when confronted with shifting monetary landscapes.

The dominant driver behind this crypto ecosystem shift involves changing rate expectations. Last week saw a stronger-than-expected United States jobs report, which added 162,000 positions to the economy. This robust employment data immediately increased odds for a Federal Reserve rate hike at the upcoming September 15 to 16 meeting. Higher government bond yields directly reduce the relative appeal of speculative investments.

Buyers now heavily price in a higher probability of tighter monetary policy ahead of critical inflation figures. The financial world eagerly awaits the United States Consumer Price Index report scheduled for Friday, September 11. This upcoming inflation print will either solidify or soften current expectations for central bank hikes. Market participants remain highly sensitive to economic metrics that might influence monetary authority decisions.

Also Read: Why a strong jobs report hit Bitcoin and Ethereum harder than the stock market

Valuation drops in the digital asset space frequently trigger severe mechanical selling. The recent cryptocurrency decline initiated a massive leverage unwinding event across the digital landscape. Exchanges recorded US$264 million in total liquidations over the last 24 hours. Positions tied to the largest blockchain accounted for US$73.33 million of this total. This liquidation volume represents a 76.73 per cent increase from the previous day.

Approximately 90 per cent of these forced sales involved long positions. This statistic indicates a complete flush of overleveraged bullish bets. Forced selling creates a dangerous feedback loop that exacerbates downward valuation momentum. Analysts must watch for stabilisation in open interest and funding rates to confirm that this leverage flush has finally run its course across major platforms.

Traditional equity venues also reflect deep participant concern regarding the broader economic outlook. Wall Street closed lower on Tuesday, September 8, 2026. Major indices surrendered substantial ground as individuals digested negative news regarding global energy supplies. The Dow Jones Industrial Average fell 628.18 points or 1.2 per cent to close at 52,786.07. The S&P 500 dropped 45.08 points or 0.6 per cent to finish at 7,673.52.

The Nasdaq Composite slipped 85.58 points or 0.3 per cent to end the session at 26,421.41. Small-cap stocks also retreated as the Russell 2000 index lost 15.44 points, or 0.5 per cent, to settle at 2,960.20. These broad declines demonstrate that equity buyers share the exact same risk-off sentiment currently gripping the digital asset space and global commodity venues.

Sector performance on Wall Street clearly illustrates the specific fears driving this equity sell-off. The Energy, Utilities, and Real Estate sectors managed to close higher despite the broader index’s decline. Conversely, Health Care, Financials, and Materials severely lagged the wider financial landscape.

The Dow Jones Industrial Average took a particularly hard hit due to a sharp pullback in healthcare. The primary catalyst for this sector rotation involves rapidly escalating geopolitical tensions involving Iran. These conflicts directly threaten global energy infrastructure and disrupt regional supply chains. Crude oil markets reacted violently to these disruptions. Brent crude briefly approached US$99.50 a barrel as attacks on regional energy facilities spooked commodity buyers. Oil eventually settled in the green as participants priced in a sustained period of elevated energy costs.

Also Read: From US$79,300 to US$82,400: Mapping the narrow corridor that decides Bitcoin’s September

Rising energy costs directly renew inflation worries among institutional and retail buyers. Crude oil nearing US$100 a barrel introduces a massive variable into future inflation calculations. Higher fuel and transportation costs inevitably filter down to consumer goods and services. This dynamic severely complicates the Federal Reserve’s mandate to maintain price stability.

The combination of strong jobs data and surging oil prices creates a perfect storm for persistent inflation. Participants now fear that the central bank might adopt an even more aggressive stance to combat these rising prices. This reality explains why both digital assets and traditional equities sold off simultaneously. Individuals simply lack the appetite to hold risk instruments when the cost of capital threatens to rise significantly soon.

The immediate outlook for the leading cryptocurrency remains sideways to bearish until the ecosystem digests upcoming inflation data. A sustained hold above US$78,000 could stabilise the asset and attract cautious buyers. A daily close below this crucial threshold would likely trigger a test of lower supports. Analysts currently target the US$76,000-US$77,600 support zone if bearish momentum continues.

Also Read: Bitcoin just broke US$81,000: The real reason is not what you think

Conversely, a cooler inflation print could allow a rebound toward the US$80,000 mark. Speculators should note that firm resistance awaits near the US$81,000 to US$82,000 range. Institutional exchange-traded fund demand continues to provide a structural bid for the asset. Macroeconomic fears completely overshadow this underlying institutional demand during the current trading week as participants await concrete economic numbers.

The combination of hawkish central bank repricing and a leveraged long squeeze has definitively tipped short-term momentum downward. The path of least resistance remains cautiously lower until the ecosystem receives concrete macroeconomic confirmation. Friday, September 11, stands out as the most critical date for near-term price discovery.

The reaction at the US$78,000 support will determine whether the financial landscape experiences a healthy pullback or a deeper correction. At the time of writing this analysis, Asian exchanges have not opened for the trading session. I eagerly anticipate analysing the Asian exchange reaction when trading begins. The opening bell in Asia will likely provide crucial clues regarding global sentiment and set the tone for the week.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post Asia has not opened yet: What will the first bell reveal about Bitcoin and oil? appeared first on e27.

Thailand’s mobility story is often told through the lens of electric vehicles, new car factories and government incentives. But at HERE Directions Bangkok 2026, the more important question was not simply what people will drive. It was how vehicles, roads, cities and public agencies will share enough intelligence to make movement safer and more efficient.

Hosted by location data company HERE Technologies at Park Hyatt Bangkok, with Amazon Web Services as co-host, the event brought together government representatives, automakers, technology firms and mobility specialists to examine the next phase of Thailand’s transport evolution. The discussion ranged from road safety and electrification to AI-assisted driving and connected urban data systems.

Also Read: Southeast Asia’s EV startups draw US$622M as clean mobility shifts from pitch to pilot

The timing matters. Thailand is already Southeast Asia’s most important automotive manufacturing base and has set an ambition for zero-emission vehicles to account for 30 per cent of total vehicle production by 2030. At the same time, Bangkok’s congestion, Thailand’s high road fatality rate and the country’s heavy dependence on motorcycles show that the mobility transition cannot be solved by swapping petrol engines for batteries alone.

Road safety is still the hardest problem

One of the clearest themes from the event was that road safety remains Thailand’s most urgent mobility challenge.

Motorcycles account for nearly half of registered vehicles in the country, making two-wheeler safety central to any national transport strategy. For millions of Thais, motorcycles are not recreational vehicles; they are the default option for commuting, food delivery, informal logistics and last-mile transport. That makes the risks harder to manage and the policy response more complex.

Location intelligence has an obvious role here. Accident hotspot alerts, road condition notifications, safer route suggestions and live traffic updates can help drivers and riders make better decisions before danger becomes unavoidable. For fleet operators, the same data can shape driver coaching, route planning and insurance risk models.

“Thailand is entering a new era of mobility where real-time decisions matter more than ever. Electrification, AI-powered driving experiences and rising expectations around road safety are transforming how people and goods move,” said Deon Newman, Senior Vice President and General Manager for Asia Pacific at HERE Technologies. “As vehicles become more connected and software-defined, location intelligence is becoming the critical foundation that helps drivers, businesses and cities make safer, smarter and more informed decisions in real time.”

The point is especially relevant in Southeast Asia, where urban transport systems are highly mixed. Cars, buses, motorcycles, tuk-tuks, delivery riders and pedestrians often share the same road space. In that environment, maps cannot be static digital replicas of roads. They need to capture changing conditions, risk patterns and local driving behaviour.

EV adoption needs more than chargers

Electrification was another major focus, but speakers treated it as part of a wider mobility shift rather than a standalone vehicle trend.

Thailand’s EV market has been expanding, supported by government incentives and investment from global and Chinese automakers. Yet the transition includes more than battery electric cars. Hybrids, plug-in hybrids, hydrogen fuel-cell vehicles, commercial fleets, electric buses and two-wheelers will all be part of the mix.

This creates a new planning burden. Drivers need to know not only where a charging station is, but whether it is available, compatible, reliable and reachable based on battery level, traffic and terrain. Fleet operators need to plan routes around charging windows and delivery schedules. Cities need to understand where infrastructure gaps are emerging.

That is where location data becomes operational rather than merely navigational. A map that can combine road networks, charging locations, energy consumption patterns and live traffic can help reduce range anxiety and improve vehicle utilisation. For logistics companies, even small gains in routing efficiency can translate into lower costs across thousands of trips.

AI turns maps into decision systems

The event also looked at how artificial intelligence is changing the role of in-vehicle navigation. Advanced driver assistance systems, or ADAS, and Navigation on Autopilot are pushing maps beyond turn-by-turn directions.

For these systems to work safely, vehicles need to understand road context: lane structures, speed restrictions, intersections, curves, construction zones and hazards ahead. AI can help interpret this environment, but it still depends on reliable underlying map and location data.

Also Read: SLEEK EV’s US$8.5M Series A funding signals a more mature EV playbook

This is where the industry is moving towards what automakers often call software-defined vehicles. In simple terms, more of the vehicle’s functions are managed and improved through software rather than fixed hardware alone. Navigation, safety alerts, driver assistance and infotainment are increasingly connected.

At HERE Directions Bangkok, neueHCT demonstrated several AI-powered driving technologies, including HCT Astra, an assisted driving platform; HCT Luna, a smart camera system; and HCT Orbis, a rider assistance system for two-wheelers. The inclusion of two-wheeler technology is notable in Thailand and the wider region, where mobility innovation often needs to start with motorcycles rather than premium cars.

Smarter cities need shared data

Beyond vehicles, the event returned repeatedly to the importance of connected data ecosystems. Smart city projects often struggle because information sits in separate systems across government agencies, transport operators, emergency services and private mobility companies.

Dr Passakon Prathombutr, Vice Chairman of iTIC and Special Expert at Thailand’s Digital Economy Promotion Agency, argued that the value of data increases when different layers can be combined. Accident data, GPS traces, road context and infrastructure information can reveal patterns that would be invisible in isolation.

“Building smarter and more sustainable cities requires more than technology. It requires the ability to connect data across agencies, infrastructure and mobility ecosystems,” he said. “When data can move seamlessly between stakeholders, this can add higher value to cities that can gain deeper insights, improve decision-making and deliver more efficient, safer and citizen-centric mobility services.”

For Bangkok, this is not an abstract ambition. The city’s transport challenges are shaped by density, legacy infrastructure, flooding risks, delivery growth and fragmented public transport options. Better data sharing could support traffic management, emergency response, public transport planning and road safety interventions.

A competitive mapping race

HERE Technologies is not alone in chasing this opportunity. Globally, it competes with Google Maps Platform, TomTom and Mapbox across mapping, location services and automotive navigation. In Southeast Asia, Grab has also built mapping capabilities to support ride-hailing, deliveries and logistics, while Waze remains influential in crowdsourced traffic information.

The competitive landscape reflects a broader shift: mapping is no longer just a consumer app category. It is becoming infrastructure for autonomous driving, urban planning, insurance, logistics, advertising, EV charging and public safety. For Thailand, that means the winning solutions will need strong local data, partnerships with public agencies and the ability to work across messy real-world conditions.

Also Read: Thailand’s AI startup push gets OpenAI backing through new public-private accelerator

The lesson from HERE Directions Bangkok 2026 is that Thailand’s mobility future will depend less on any single technology than on how well different systems talk to one another. EVs, AI-assisted driving and smart city platforms may capture the headlines, but their impact will be limited if roads, vehicles and institutions continue operating on incomplete information.

In Southeast Asia, where mobility is crowded, informal and fast-changing, the next breakthrough may not look like a futuristic car. It may be a better decision made a few seconds earlier — by a driver, a fleet manager, a traffic controller or a city planner.

The post Thailand’s mobility future will be decided by data, not just vehicles appeared first on e27.

For years, the semiconductor race was largely about making transistors smaller. That contest is far from over, but the AI boom has shifted part of the battleground elsewhere: how multiple chips are assembled, connected and kept reliable inside a single package.

That is the problem Synopsys and Singapore’s Agency for Science, Technology and Research (A*STAR) are now trying to address.

Also Read: The factories are coming. Southeast Asia’s real race is to build what surrounds them

The US-based chip design software company and Singapore’s national research agency have signed a memorandum of understanding to jointly develop advanced semiconductor-packaging and simulation technologies for artificial intelligence and high-performance computing.

The collaboration will focus on advanced packaging and chiplet-based designs. Chiplets are smaller specialised chips that can be combined in one package to function like a larger, more powerful system. Instead of relying on one monolithic chip to do everything, companies can mix and match computing, memory, networking and other functions in a more modular way.

This approach is increasingly important for AI and high-performance computing, where systems need far more processing power, memory bandwidth and energy efficiency than traditional chip designs can easily deliver. But it also creates new engineering challenges. When several chips are packed tightly together, heat, stress, warping and material behaviour become harder to predict.

Synopsys said digital modelling can help assess how a package is likely to perform before companies spend time and money on physical prototypes. In an industry where development cycles are long and fabrication mistakes are expensive, being able to simulate reliability early can make a meaningful difference.

Why packaging now matters as much as design

The collaboration will run through the ASTAR IME-Ansys Joint Innovation Consortium for Semiconductor Excellence, involving ASTAR’s Institute of Microelectronics and Ansys, which is now part of Synopsys. The consortium will act as a platform for companies and researchers to conduct joint research on advanced System-in-Package technologies.

System-in-Package, or SiP, refers to the integration of multiple chips or components inside one package. It is already used in areas such as smartphones, automotive electronics and wearables, but AI computing is pushing the technology to far greater levels of complexity.

The initial phase of the consortium’s work will focus on mechanical design, modelling and analysis. The partners aim to examine issues including package and wafer warping, thermo-mechanical stress, solder-joint reliability and moisture-induced failures in multi-chiplet designs.

Also Read: Southeast Asia’s chip-hub ambition is colliding with its chip-smuggling problem

These may sound like back-end engineering details, but they are central to whether future AI systems can be manufactured at scale. A high-performance chip package may fail if it bends during production, develops microscopic cracks under heat, or suffers from unreliable solder joints over time. For AI data centres, autonomous systems, advanced manufacturing and next-generation consumer electronics, reliability is not optional.

ASTAR IME will lead the consortium’s research, drawing on its semiconductor packaging capabilities and research platforms. Synopsys will provide trial licences to its Ansys simulation software to ASTAR IME and up to 10 member companies. The consortium also plans to explore projects with universities and offer technical training.

That training element is particularly relevant for Singapore and the wider region. Semiconductor ecosystems are not built only on fabs and equipment; they also require engineers who understand materials, design, thermal behaviour, electronics, manufacturing constraints and simulation tools.

Singapore’s semiconductor bet

Singapore has long played an outsized role in the global semiconductor supply chain. It is home to wafer fabrication, assembly and test operations, equipment suppliers, materials companies and regional headquarters for multinational chip firms. While it does not compete with Taiwan or South Korea in leading-edge logic manufacturing, it has positioned itself as a serious hub for specialty chips, advanced packaging, research and manufacturing services.

That positioning matters as the global chip industry becomes more geopolitically fragmented. The United States, China, Europe, Japan, South Korea and Taiwan are all investing heavily in semiconductor capabilities, driven by AI demand and concerns over supply chain resilience. Southeast Asia, meanwhile, has become more important as companies diversify manufacturing footprints and look for politically stable, technically capable locations.

Malaysia is already a major assembly and testing hub, particularly in Penang. Vietnam is attracting interest in chip design and back-end manufacturing. Thailand and the Philippines have existing electronics manufacturing bases. Singapore’s advantage lies in its research institutions, talent base, intellectual-property protections and proximity to both global companies and regional manufacturing networks.

The Synopsys-A*STAR partnership fits neatly into that strategy. Rather than trying to win every part of the chip supply chain, Singapore has been focusing on areas where deep engineering, industry collaboration and applied research can create defensible value.

Terence Gan, Executive Director of A*STAR IME, said advanced packaging is becoming a critical differentiator in chip innovation as systems grow more complex and AI-driven. He added that the collaboration can reinforce Singapore’s position as a semiconductor innovation hub.

Simulation becomes a strategic layer

For Synopsys, the tie-up also reflects how electronic design automation companies are expanding beyond traditional chip design software. The company’s acquisition of Ansys brought simulation capabilities closer to chip and system design, at a time when the boundaries between semiconductor design, packaging and system-level engineering are blurring.

In AI hardware, performance is no longer determined only by the processor. Memory access, interconnects, power delivery, cooling and package architecture all shape the final system. This makes simulation more strategic. Companies need to understand how a design will behave electrically, mechanically and thermally before it reaches production.

Sukhwan Moon, Synopsys Vice President of Asia-Pacific Sales, said the collaboration can support local companies in reliability, performance and scalability, while helping speed time-to-market for AI and high-performance computing technologies.

That time-to-market pressure is intense. AI infrastructure demand has created a rush for more powerful chips, faster networking and more efficient computing systems. Cloud providers, hyperscalers, chip startups and electronics manufacturers are all trying to move quickly, but the hardware cycle remains unforgiving. Mistakes in design or packaging can delay products by months.

The competitive landscape

Synopsys operates in a highly concentrated but fiercely competitive market. Its main global rivals include Cadence Design Systems and Siemens EDA, which also provide chip design, verification and electronic design automation tools. In simulation and engineering software, the enlarged Synopsys now overlaps with companies such as Keysight Technologies, Altair and Dassault Systèmes in certain areas, depending on the application.

Also Read: AI demand lifts Malaysia’s chip sector, but not every player wins

In advanced packaging, competition is not limited to software. Foundries, outsourced semiconductor assembly and test players, and integrated device manufacturers are all building capabilities around chiplets, 2.5D and 3D packaging. TSMC, Intel, Samsung, ASE, Amkor and JCET are among the companies shaping this market globally. For Southeast Asia, this creates both opportunity and pressure: the region can capture more value, but only if it moves beyond low-cost manufacturing into higher-value engineering and research.

A small MoU in a much larger race

The MoU between Synopsys and A*STAR is not a chip factory announcement, nor does it come with the headline-grabbing capital expenditure often associated with semiconductor projects. Its importance lies elsewhere.

Advanced packaging is becoming one of the key ways the chip industry keeps improving performance as traditional scaling becomes harder and more expensive. For AI, where demand for computing power continues to surge, the ability to design reliable multi-chip systems could determine which companies and countries capture the next wave of value.

For Singapore, the partnership strengthens a role it has been cultivating for years: a neutral, research-driven and industry-connected node in the global semiconductor network. For Southeast Asia, it is another sign that the region’s chip opportunity is widening beyond assembly lines.

The future of AI hardware will not be decided only in data centres or wafer fabs. It will also be shaped in the less visible world of packaging labs, simulation platforms and reliability testing — precisely where this collaboration intends to operate.

The post Synopsys, A*STAR team up to tackle AI chip packaging challenges appeared first on e27.

Artificial intelligence has dramatically changed the way startups are built.

Today, a founder can describe a product idea, open a tool such as Claude or OpenAI, generate hundreds of lines of code, build a prototype and present it as an “AI-powered innovation” within days. What once required a technical team, months of development and significant capital can now be achieved remarkably quickly.

The barrier to building software has never been lower.

But as the ability to build accelerates, another challenge becomes increasingly important: validation.

AI can generate code, analyse information, summarise complex material and produce remarkably convincing answers. But a convincing answer is not necessarily a correct one. And when AI-generated outputs move from a demo environment into industries where mistakes have real consequences, the difference between “working” and “working reliably” becomes critical.

Healthcare is perhaps the clearest example.

For a healthcare AI startup, the question is not simply whether a model can produce an answer. It is whether that answer is clinically reliable, generated from appropriate data, reproducible across relevant populations and settings, understandable to the intended user, and safe enough to inform a clinical decision.

This is where human-in-the-loop verification becomes more than a safety feature. It becomes part of the product itself.

The clinician is not the bottleneck

There is sometimes an assumption that AI automation becomes more valuable as humans are removed from the workflow. In healthcare, that assumption can be dangerous.

AI can process enormous volumes of information far faster than a human. It can identify patterns across patient records, compare information against large knowledge bases and surface potentially relevant findings. But it does not automatically understand the complete clinical context in which those findings will be used.

A clinician does. That is why the most useful healthcare AI may not be the system that attempts to replace clinical judgement, but the one that augments it.

Regulation is increasingly reflecting this distinction. The US Food and Drug Administration’s January 2026 final guidance on Clinical Decision Support Software clarifies the criteria for certain clinical decision-support functions to qualify as non-device software. One important criterion is whether the software enables healthcare professionals to independently review the basis of its recommendations rather than primarily relying on the software’s output.

Singapore’s Health Sciences Authority has similarly refined its framework for Software as a Medical Device (SaMD) and Clinical Decision Support Software (CDSS), as outlined in its update on SaMD risk classification and CDSS qualification guidelines. Its July 2025 revision added, among other changes, a criterion concerning whether CDSS recommendations are based solely on established clinical guidelines when determining whether software qualifies as a non-medical device.

The message for founders is important: automation does not automatically mean removing the professional from the loop.

Depending on its intended purpose, functionality and risk, software may fall within medical-device regulation or qualify for a non-medical-device pathway. Either way, the product needs to be designed around clearly defined accountability.

The clinician interprets the recommendation, considers the patient’s circumstances and decides whether to accept, modify or reject it.

Also Read: Vietnam’s healthtech boom has a talent problem nobody is talking about

“Accurate” is not enough

AI hallucination is often discussed as a technical problem. In healthcare, it is a product and safety problem.

A generative AI system can produce an answer that is fluent, structured and persuasive while being completely wrong. A fabricated reference, incorrect interpretation of a medical record or inappropriate recommendation could have consequences far beyond a poor user experience.

This means healthcare AI cannot be evaluated simply by asking: “How accurate is the model?”

The more useful questions are: Accurate for whom? Under what conditions? Compared with what reference standard? Using which data? And in which real-world population?

Traditional metrics such as sensitivity, specificity, precision, recall and area under the receiver operating characteristic curve (AUC) remain valuable. But a strong metric on a controlled dataset does not automatically translate into reliable performance in clinical practice.

A model can perform exceptionally well in one dataset and behave differently when exposed to another hospital, patient population, imaging device, documentation style or clinical workflow.

This is why validation must extend beyond the model itself.

It includes the quality and representativeness of the data, external validation, clinical workflows, human factors, usability, monitoring and performance after deployment. For regulated software, lifecycle management, verification and validation, change management and post-market considerations are increasingly important parts of the development process. HSA, for example, maintains a lifecycle-oriented framework for software medical devices alongside its SaMD and CDSS classification guidance.

The new startup moat may be trust

For founders, this creates an important strategic shift.

The competitive advantage of an AI startup may no longer be simply how quickly it can build a model.

If thousands of startups can use the same foundation models and increasingly powerful coding tools, the ability to produce a prototype becomes less differentiated.

The harder question becomes: Can you prove that what you built works?

That proof may become one of the strongest forms of competitive advantage.

A startup that combines AI automation with genuine domain expertise, structured validation, transparent outputs and continuous monitoring can build something considerably more defensible than a product that simply places a large language model on top of an existing workflow.

This is particularly relevant for founders entering regulated or high-stakes industries. Domain experts should not be brought in merely to satisfy an advisory requirement after the product has been built. Their expertise should influence the product architecture, validation strategy, workflow design and definition of failure.

Also Read: Healthtech in South and Southeast Asia – Seeing beyond the “obvious”

Human-in-the-loop should therefore not be viewed as a limitation on AI.

It is a mechanism for making AI deployable.

The best systems will know what to automate, when to request human verification and, critically, when not to provide an answer at all.

From “AI versus humans” to “AI plus humans”

The future of AI in healthcare is unlikely to be a simple contest between artificial intelligence and human intelligence.

It is more likely to be a carefully designed partnership.

AI brings scale, speed and the ability to process enormous amounts of information. Humans bring contextual understanding, professional judgement, ethical responsibility and the ability to challenge an output when something does not look right.

The real innovation lies in designing the interface between the two.

As AI lowers the cost and time required to build software, founders will increasingly be judged on something beyond how quickly they can produce a demo.

They will be judged on whether they can demonstrate that their product works, understand where it can fail, and build the mechanisms to detect and manage those failures.

In an age where almost anyone can build software with a prompt, building is becoming easier. Proving is becoming harder.

And for high-stakes AI, that may be where the real startup advantage lies.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post The missing layer in AI innovation: Human verification appeared first on e27.

Every hyperscaler courting Southeast Asia now performs the same reassurance ritual. Ask Microsoft, Google, or AWS about the environmental cost of the data centres they are racing to build across the region, and the answer arrives pre-packaged: efficient cooling, renewable offsets, and community engagement.

Worse, tech giants have even started telling reporters that their facilities are “less thirsty” than before, part of a broader industry effort to get ahead of mounting public anger in the US over how much water AI infrastructure consumes.

Also Read: The AI server boom in Southeast Asia: Why data centres are running out of power

That messaging has not yet reached Gelang Patah, a town in Johor, Malaysia, where residents gathered outside a data centre construction site earlier this year holding a straightforward complaint: there was not going to be enough water left for them. It is a small protest that points to a large problem, and Southeast Asia’s AI boosters would rather not dwell on it.

Johor is the test case, and it is already straining

Johor is Southeast Asia’s fastest-growing data centre hub for a reason that has nothing to do with Malaysia’s own digital ambitions. When Singapore froze new data centre approvals between 2019 and 2022 to protect its limited land and water resources, global operators simply moved their plans across the causeway. Johor’s aggregate capacity has since surged towards 5.8 gigawatts, and the state now hosts dozens of operational facilities feeding off a water and power system that was never built for this scale of industrial demand.

The numbers are no longer abstract. A single 100-megawatt facility can consume in the region of 4 million litres of water a day for cooling alone; Johor officials have described hyperscale sites drawing roughly 200 times more water than an ordinary industrial user.

Regulators have responded by rejecting a meaningful share of new applications, raising industrial water tariffs, and telling the largest prospective tenants to wait until at least mid-2027 for guaranteed water and power connections. One estimate puts committed demand from the state’s data centre pipeline at more than 800 million litres a day, against roughly 140 million litres of infrastructure actually capable of delivering it today.

Malaysia’s federal government has all but admitted the model is unsustainable in its current form. Prime Minister Anwar Ibrahim told parliament this year that Malaysia has stopped approving new non-AI-linked data centres altogether, betting that AI-branded projects still qualify for green-lighting even as electricity demand from the sector is projected to climb to nearly a third of the country’s entire power supply by 2035, up from around seven per cent today.

Malaysia is not the exception; it is the preview

Treat Johor as a warning rather than an isolated case, because the same arithmetic is playing out across the region with fewer headlines. Indonesia’s Java-Bali grid, which carries most of the country’s data centre load, was already running near capacity before a single new AI facility came online, and the grid remains roughly 60 to 65 per cent dependent on coal, a fact that sits awkwardly against hyperscalers’ net-zero pledges. Thailand’s data centre power demand reportedly grew fourfold between 2020 and 2024 while generation capacity crept up by less than a tenth of that.

Vietnam, meanwhile, is attracting hyperscale investment on the strength of cheap land and labour even as parts of the country face weekly power cuts during summer peaks.

Also Read: Breaking into the data centre sector: Beyond technical expertise

A recent Bain and Standard Chartered analysis framed this plainly: the binding constraint on Southeast Asia’s AI-driven growth is not capital or ambition, it is the grid itself, and the region’s transmission and distribution networks have not kept pace with the concentrated, high-value demand that data centres represent. Roughly 35 to 45 terawatt-hours of incremental demand is expected across the region’s hubs by 2030 (Singapore, Johor, Bangkok, Greater Jakarta, Manila, and Batam) landing on infrastructure largely designed for a slower, more distributed pattern of growth.

The speculative capacity problem makes this worse

What makes the resource strain harder to justify is that a meaningful share of the demand driving it may not even be real yet. Malaysia’s Energy Commission has found that data centres were drawing less than half of their declared maximum electricity demand as of mid-2025, prompting officials to flag the likelihood of speculative applications — developers reserving grid capacity and water allocations well ahead of actual tenant commitments, effectively queue-jumping scarce resources against uncertain future need.

That is a familiar pattern from past infrastructure bubbles, and it means some of the water and power tension communities are living with today is being generated by capacity that may never be fully utilised.

This is precisely why Johor’s response, however belated, is worth taking seriously as a template rather than dismissing as friction. The state has rejected close to a third of data centre applications over sustainability shortfalls, mandated a shift towards reclaimed wastewater instead of municipal supply for new approvals, and built a dedicated water reuse programme aimed squarely at the industry.

None of that has fully closed the gap between committed capacity and available infrastructure. But it is a materially more honest starting point than the alternative most of the region has defaulted to: approve first, measure the damage later.

Southeast Asia should not import a problem it can still design around

The uncomfortable truth is that Southeast Asia has a genuine opportunity here that most of the world does not. Its AI data centre boom is still young enough that grid interconnection, water accounting standards and siting rules can be built deliberately, rather than retrofitted after the fact the way the United States is now attempting. The ASEAN Power Grid interconnection and philanthropic clean-energy pledges for surrounding communities are steps in the right direction, but they remain medium-term fixes for a strain that is already showing up in tariffs, deferred approvals and community protest today.

Also Read: The US$5 trillion AI data-centre buildout unleashes the paradox that limits its returns

Governments across the region would do well to stop treating data centre investment announcements as unambiguous economic wins and start asking the harder question Johor is now being forced to confront: what does this facility cost the people living next to it, in water, in power, and in a grid that other industries and households also depend on?

The alternative is a region that spent its AI infrastructure boom exporting the same environmental trade-offs Silicon Valley is only now starting to reckon with — except this time, with far less capacity to say no.

The post SEA’s AI boom has a water problem it cannot offset away appeared first on e27.

Southeast Asia’s e-commerce story is no longer just about search bars, discount vouchers and marketplace rankings. Increasingly, shoppers are discovering what to buy through livestreams, short videos, creator reviews and affiliate content. And that shift is now large enough to reshape the region’s online retail economy.

Content commerce gross merchandise value across Shopee, TikTok Shop, and Lazada reached US$49.7 billion in 2025, almost doubling from the previous year, according to Momentum Works’s latest report, Live Commerce in Southeast Asia 2026. The Singapore-headquartered research and venture outfit estimates that US$33.8 billion was transacted in the first half of 2026 alone. If the current pace holds, the segment is on track to hit US$77.9 billion for the full year.

Also Read: The rise of live commerce in Asia and adoption of BeLive by retailers

The more telling number is not just the headline GMV, but content commerce’s share of the region’s e-commerce mix. In the first half of 2026, it accounted for 37 per cent of Southeast Asia’s platform e-commerce GMV, up from 20 per cent in 2024. In other words, what was recently treated as an add-on marketing channel has become a major sales engine.

For founders, brands and marketplace operators in the region, this marks an important turning point. Southeast Asia’s e-commerce markets — from Indonesia and Thailand to Vietnam, the Philippines, Malaysia, and Singapore — have long been shaped by mobile-first behaviour, price sensitivity, and high social media usage. Content commerce sits at the intersection of all three. It makes shopping more entertaining, but also more immediate: a product demo, a creator recommendation, and a checkout button can now sit within the same customer journey.

Live commerce moves into the operating core

Live commerce has been the most visible part of this shift. The format allows sellers, creators, and brands to demonstrate products in real time, answer questions, and trigger purchases through limited-time offers or platform vouchers. In categories such as beauty, fashion, household goods, and fast-moving consumer products, it has become a daily operating channel rather than a campaign experiment.

Momentum Works notes that for many brands, the question is no longer whether they should go live, but what role live should play in the broader business. Some use it mainly for conversion, pushing volume during platform sales days. Others use it to educate consumers on new products, build trust in unfamiliar brands, or move slower-selling inventory.

That distinction matters because live commerce does not work equally well for every product. A low-priced lipstick, snack bundle, or kitchen gadget can benefit from quick demonstrations and impulse buying. Higher-consideration purchases may need more education, reviews, and repeat exposure before a customer checks out. Execution quality also matters: the host, script, pacing, product assortment and incentives can materially affect sales.

For now, the returns from live remain attractive for many operators. But Momentum Works argues that these returns are unlikely to stay unusually high forever. They are being supported by growing consumer attention, platform incentives and a competitive environment that is still maturing. As more brands, agencies, sellers and creators develop similar capabilities, the cost of standing out will rise.

The next battleground: brandformance

This is where “brandformance” enters the conversation. The term, a blend of brand building and performance marketing, captures a problem many e-commerce teams face: short-term conversion can be measured instantly, but long-term consumer preference is harder to track.

Live commerce is naturally performance-driven. A seller can see how many viewers joined, how long they stayed, which products were clicked and what was purchased. That makes it appealing in a region where marketing budgets are often tied closely to measurable outcomes. But if every brand is running live sessions with similar scripts, discounts and affiliate networks, performance alone becomes easier to copy.

Also Read: Elevating your e-commerce strategies with livestreaming and hero products

The longer-term advantage may sit with companies that use content not only to sell, but to build memory and trust. That could mean explaining why a skincare product works for humid climates, why a halal-certified food product matters to Muslim consumers, or how an electronics brand supports after-sales service in provincial cities. In fragmented Southeast Asian markets, where language, culture, logistics, and purchasing power differ widely, local relevance is not a minor detail.

The challenge is that many brands still treat live commerce as a standalone sales machine. It is visible, measurable and relatively easy to justify internally. Short videos, affiliate reviews and community content can be harder to attribute, even when they play a crucial role in creating demand before the livestream begins.

AI lowers the cost of execution

The report also points to a structural change that could compress the advantage of skilled operators: AI live. In selected cases, Momentum Works says AI-driven live operations cost around 20-25 per cent of a comparable human setup while achieving around 80 per cent of human livestream GMV per hour on average.

That has significant implications. Capabilities that once took agencies, brands, and livestream studios years to build (scripting, scheduling, product explanations, host consistency, and basic audience interaction) are becoming more accessible through technology and platform tools. For smaller sellers, this could lower the barrier to entry. For larger brands, it could reduce operating costs and allow more always-on content.

But it also creates a strategic problem. If everyone can access similar tools, operational capability alone becomes less defensible. The differentiator shifts to what cannot be automated as easily: product quality, customer insight, creative direction, creator relationships, community trust and brand positioning.

This matters in Southeast Asia because the region’s ecommerce growth has often been fuelled by intense marketplace competition and subsidised demand. As subsidies normalise and consumer acquisition becomes more expensive, brands will need more than efficient livestream operations. They will need reasons for shoppers to return without being pulled only by the next discount.

China offers lessons, not a template

China remains the global reference point for live commerce. Its ecosystem is more mature, with advanced livestream infrastructure, professional creator networks, high-frequency shopping behaviour and deeper use of data and automation. Southeast Asian platforms, brands and sellers have borrowed heavily from that playbook.

Yet Momentum Works cautions that China should be seen as a map, not a blueprint. Southeast Asia is not one market. Creator economics in Indonesia differ from Singapore. Consumer trust patterns in Vietnam may not mirror those in Thailand. Payment habits, logistics reliability, local languages and platform dynamics vary sharply across the region.

That means the next phase of content commerce will likely be less about copying a single model and more about adapting formats market by market. A livestream strategy that works in Bangkok may need to be rebuilt for Manila. A short-video approach that drives discovery in Jakarta may not translate neatly to Ho Chi Minh City.

Also Read: How AI, AR, and live streaming are changing the online shopping experience

The broader lesson is clear: live commerce has become infrastructure, but it is not the entire content commerce strategy. As the channel matures, brands that over-invest in live while under-funding short video, affiliates and review-led discovery risk mistaking the checkout moment for the whole customer journey.

For Southeast Asia’s digital economy, the US$77.9 billion forecast is a sign of how quickly shopping behaviour is changing. The next question is not whether content will shape e-commerce, but who can turn attention into durable customer relationships once the easy growth fades.

The post Southeast Asia’s live commerce boom enters its harder second act appeared first on e27.

Capital A’s (formerly AirAsia Group) latest numbers tell a company coming out of crisis, but not yet one firing evenly across all engines.

The Malaysia-based group, which has spent the past few years restructuring after the pandemic and disposing of its airline business, reported second-quarter revenue of about US$193 million, up 9 per cent year-on-year. For the first half of 2026, revenue stood at about US$376 million, a 4 per cent increase from a year earlier.

On the surface, that points to stability. Capital A also reported profit after tax of about US$6 million for the quarter and US$11.9 million for the first half, giving it another profitable quarter after the airline disposal.

Also Read: AirAsia unit Teleport buys stake in Indonesia’s ‘Uber for logistics’ Kargo Technologies

But the recovery is more uneven than the topline implies. Growth is being driven mainly by two units: Asia Digital Engineering (the aircraft maintenance, repair, and overhaul business) and Teleport (the logistics arm). Together, ADE and Teleport accounted for more than 70 per cent of first-half group revenue.

That leaves the rest of the portfolio (AirAsia MOVE, AirAsia Next, and Santan) with a harder job to prove that Capital A can build a broad-based, asset-light aviation services and digital platform business beyond the airline brand that made it famous.

Profitability returns, but margins remain thin

Capital A’s return to profitability is meaningful. The group has exited PN17 status, a classification for financially distressed companies on Bursa Malaysia, and is trying to rebuild investor confidence around a cleaner corporate structure.

Yet the profit margin leaves little room for error. Second-quarter profit after tax of around US$6 million on revenue of US$193 million implies a net margin of roughly 3.1 per cent. For the first half, profit after tax of US$11.9 million on US$376 million revenue works out to about 3.2 per cent.

For a group still in transition, that is not alarming by itself. But it does mean the turnaround remains vulnerable to foreign exchange movements, interest costs, lease obligations, capital expenditure and slower volumes.

The operating picture is also less flattering. First-half net operating profit fell 21 per cent year-on-year to about US$16 million, despite revenue growth. Capital A said core group net operating profit rose 6 per cent after adjusting for the loss of aviation interest income following the airline disposal.

That adjustment may be fair, but it is also doing a lot of work. The reported number shows operating profit declined. The adjusted number supports the recovery story. Investors will want a clearer bridge between the two.

Group EBITDA also fell 5 per cent in the first half, even as revenue rose 4 per cent. That suggests either costs are rising faster than sales, or the revenue mix is tilting towards lower-margin activities.

ADE and Teleport carry the group

The strongest part of the update is ADE. The aircraft maintenance unit reported second-quarter revenue of about US$67.6 million, up 29 per cent year-on-year, with EBITDA of about US$16.4 million. Capital A said hangar slots are booked through next year and that ADE is building a new four-line maintenance hangar.

That demand backdrop is credible. Southeast Asia’s airline industry is still rebuilding capacity after the pandemic, while narrowbody aircraft fleets across the region need maintenance as utilisation rises. Supply-chain delays and aircraft delivery bottlenecks have also made maintenance capacity more valuable.

The question is how much cash ADE will need to keep growing. Maintenance is not a pure software-style business. Tools, hangars, engineering talent and certifications require investment, and depreciation will rise as capacity expands. EBITDA may look healthy while free cash flow tells a more complicated story.

Teleport also showed momentum. Second-quarter revenue rose 22 per cent year-on-year to about US$74 million, while first-half revenue grew 21 per cent to about US$147.6 million. Tonnage in the quarter reached 85,877 tonnes, up 11 per cent year-on-year, and parcel volume jumped 79 per cent to 56.6 million.

Also Read: AirAsia aims to fulfill super app ambition with upcoming launch of ride-hailing services in Malaysia

For a logistics business operating in a softer global freight market, that is a solid result. But margins remain modest. Teleport’s second-quarter net operating profit was about US$1.8 million on US$74 million in revenue, implying an operating margin of around 2.4 per cent. Profit after tax was around US$1 million.

There is also some selective framing. First-half tonnage was 182,660 tonnes, which means first-quarter tonnage was around 96,783 tonnes. On that basis, second-quarter tonnage declined sequentially even as the year-on-year comparison looked positive.

Consumer units still have work to do

AirAsia MOVE, Capital A’s travel platform, is where the pressure is more visible. The unit reported second-quarter revenue of about US$22.9 million, up 5 per cent year-on-year. But its WANO B2B business contributed 11 per cent of total revenue, or roughly US$2.5 million.

Excluding WANO, MOVE’s underlying revenue appears to have declined year-on-year. Flight sales also fell 3 per cent, which Capital A attributed to an 11 per cent reduction in AirAsia capacity. That explanation is reasonable, but it underlines MOVE’s continued dependence on the AirAsia airline ecosystem.

AirAsia Next, which includes loyalty and licensing activities, remains profitable. It posted second-quarter revenue of about US$18.6 million, EBITDA of US$6.2 million and net operating profit of US$5.2 million. But part of the growth came from non-aviation licensing fees and AirAsia Rewards, where revenue recognition can be influenced by points redemptions. The company said redemptions rose 34 per cent, helping revenue but also increasing redemption expenses.

Santan, the group’s food business, remains small. Second-quarter revenue was about US$10.7 million, broadly flat on a normalised basis, while passenger volume fell 14 per cent due to airline capacity constraints. Its push into e-commerce through TikTok and Shopee is sensible, but Capital A did not disclose the absolute revenue base, making the 30 per cent quarter-on-quarter growth figure hard to assess.

Rivals are not standing still

Capital A’s challenge is that each part of the group competes with specialised players. ADE faces established maintenance providers such as SIA Engineering, ST Engineering Aerospace, GMF AeroAsia and Lufthansa Technik Philippines. Teleport competes in a crowded logistics market against DHL, FedEx, UPS, J&T Express, Ninja Van and regional cargo operators. AirAsia MOVE is up against Traveloka, Agoda, Booking.com, Trip.com and airline direct channels. That makes execution harder: Capital A is not fighting one market battle, but several at once.

The balance sheet update also leaves questions unanswered. Capital A said shareholders’ equity is comfortably above US$119 million and operating cash flow was about US$35.7 million. It also said refinancing reduced interest expenses.

Those are positive signs. But without clearer disclosure on total debt, net debt, lease liabilities, cash balance, capital expenditure commitments and interest coverage, it is difficult to judge how strong the balance sheet really is.

Also Read: AirAsia calls off US$10M acquisition of Gojek Thailand’s fintech arm: report

The fairest reading is that Capital A is in better shape than it was during the depths of its restructuring. ADE and Teleport are growing, the group is profitable again, and the PN17 overhang has been removed.

But this is not yet a broad, high-margin recovery. It is a narrower turnaround led by two operating units, while consumer-facing businesses remain tied to airline capacity, accounting-sensitive revenue streams and early e-commerce bets. Capital A has stabilised. Now it has to prove the new group can compound.

The post Teleport powers Capital A’s rebound, but thin margins show logistics remains a hard road appeared first on e27.

For many consumer app companies, the hardest part is no longer building the first product. It is finding enough growth capital to keep buying users profitably once a product has already shown traction.

Tevo, a consumer and AI apps company from Southeast Asia, is now turning to a financing model designed for exactly that gap.

The company has secured US$10 million in non-dilutive user-acquisition financing from PvX, a Singapore-based platform that provides growth capital for mobile gaming and consumer app businesses.

Also Read: Tevo secures seed funding, strikes partnership with Vietnam’s MobiFone

Tevo said the facility will be used to scale user acquisition across its portfolio of consumer apps and AI products.

Unlike a traditional equity round, non-dilutive financing does not require the company to sell shares. In the app economy, this form of capital is often tied to marketing performance: companies use the funds to acquire users and then repay the financing from revenues generated by those users. For founders, the appeal is straightforward. If the unit economics already work, they can spend more on growth without giving up ownership.

Tevo operates around 45 consumer and AI apps across work utilities, education and entertainment. Collectively, those products have generated nearly 200 million installs globally, according to the company. The new capital will go into priority markets, higher marketing spend on user cohorts that have already proven profitable, and further investment in AI-native product features.

“Non-dilutive UA financing lets us put more capital behind products that already have product-market fit,” said Thanh Luu, CEO and founder of Tevo. “This facility also gives us the flexibility to scale globally and accelerate Tevo’s 2030 vision as the leading company in AI apps and services, and among the top five largest mobile apps and games companies from Southeast Asia.”

Why user acquisition financing is gaining ground

Tevo’s deal points to a wider change in how consumer app companies are funding growth. For years, many app businesses relied on venture capital to finance user acquisition, even when the money was being spent on paid marketing rather than product development or hiring. That made sense during the low-interest-rate era, when investors were willing to fund aggressive growth. But the downturn in venture funding has forced founders to think more carefully about what kind of capital fits each use case.

User acquisition is a different problem from building a new product. If a company has enough data to show that a customer acquired for US$1 can eventually generate more than that in revenue, then the risk profile becomes more measurable. In that case, performance-linked financing can be a better fit than equity capital, particularly for founders who do not want to dilute themselves just to increase ad spend.

PvX said it has surpassed US$750 million in committed user acquisition financing for mobile gaming and consumer app companies globally. Its focus on Singapore as a base is also notable. Southeast Asia has produced major gaming and consumer internet companies, but the region still has relatively few scaled consumer app platforms with global reach. Financing models such as PvX’s could help bridge that gap by giving app operators access to capital based on revenue performance rather than venture-market sentiment.

Also Read: PvX lands MIT investment to fund the next wave of app user acquisition

The timing is also important. Artificial intelligence has lowered the barrier to launching new consumer software products, from study tools and productivity assistants to content and entertainment apps. But it has not solved the distribution problem. App stores are crowded, advertising costs can rise quickly, and winning users requires both data discipline and capital. Companies that already run multiple apps have an advantage because they can test, optimise, and redeploy learnings across a portfolio.

Southeast Asia’s consumer app opportunity

Southeast Asia is a mobile-first region, with large young populations, high social media usage, and deep familiarity with digital services. Yet many of the world’s biggest consumer app companies still come from the US, China, Europe, Turkey, Israel and India. Southeast Asian startups have built strong positions in ride-hailing, e-commerce, fintech and gaming, but fewer have become global consumer app factories.

That is what makes Tevo’s positioning interesting. Rather than focusing on one flagship app, the company runs a broad portfolio across practical and entertainment-led categories. Work utilities and education apps can offer recurring use cases, while entertainment products can scale quickly if they find the right audience. AI adds another layer, allowing companies to turn common consumer needs — writing, studying, editing, searching, creating — into lightweight software experiences.

The challenge is that portfolio app businesses live and die by execution. Downloads alone do not guarantee long-term value. Retention, monetisation, ad efficiency, subscription conversion, and churn matter more than headline install numbers. Tevo’s nearly 200 million installs give it a base to build from, but the real test will be whether additional user acquisition spending can produce users who stay and pay.

This is where non-dilutive capital can be both useful and unforgiving. It rewards companies with strong data and clear payback periods, but it also exposes weak assumptions quickly. If marketing spend is pushed into channels or countries where users do not convert, the model breaks down. For Tevo, the stated focus on “proven cohorts” suggests the company intends to put capital behind segments where performance is already visible.

The competitive field

Tevo is not alone in chasing the consumer AI apps opportunity. Globally, it sits in a competitive field that includes portfolio app operators such as Turkey’s HubX, which runs more than 40 mobile apps across AI, education, health and fitness and has surpassed 600 million downloads. HubX recently announced an investment of up to US$75 million from Point72 Investments at a US$1.2 billion pre-money valuation, making it Turkey’s first consumer-apps unicorn.

Other global rivals include mobile-first app studios and subscription app companies building AI tools for productivity, learning, photo editing, wellness and entertainment. In Southeast Asia, the field is less crowded at scale, but local gaming studios, AI productivity startups, and consumer internet firms could all move into overlapping categories as AI app demand grows.

That competitive pressure makes distribution capital more important. AI features can be copied quickly, and app store rankings are volatile. Companies that understand paid acquisition, localisation, monetisation, and rapid product iteration are more likely to survive than those relying only on novelty.

Also Read: PvX bags US$10.5M as cohort financing goes mainstream

For Southeast Asia, Tevo’s financing is also a sign that regional consumer app companies are beginning to access the same specialised capital structures used by more mature app markets. If the company can translate financing into sustainable global growth, it could help widen the region’s startup narrative beyond marketplaces, fintech, logistics and enterprise SaaS.

The US$10 million facility is not a traditional funding round, and it does not carry the signalling effect of a headline valuation. But that may be the point. In a market where founders are being pushed to grow more efficiently, capital that follows performance — rather than hype — may become a more common way for consumer app companies to scale.

The post Tevo secures US$10M from PvX to scale its consumer AI apps without selling equity appeared first on e27.

In April 2026, the United States announced 83,387 job cuts. 26 per cent of them named artificial intelligence as the reason, the second consecutive month that AI was the top cited cause. Behind those numbers is a quieter story that is going to shape an entire generation of careers.

Stanford economists Erik Brynjolfsson, Bharat Chandar, and Ruyu Chen released findings showing that employment among 22 to 25 year olds in AI-exposed jobs has dropped between 16 and 20 per cent in software development as the trend accelerated into 2026. Older workers in the same roles are largely untouched. The cut is happening at the bottom of the ladder, not the top.

Universities and bootcamps have responded the way they usually do. Add more AI to the curriculum. Bachelor’s level AI programmes in the US grew 114 per cent from 2024 to 2025, jumping from 90 to 193 programmes. New AI majors are launching at Northwestern, Carnegie Mellon, and dozens of other universities. Coding bootcamps now market AI tracks, prompt engineering modules, and LLM integration certificates. The reflex is consistent. If AI is reshaping work, teach more AI.

This reflex is producing graduates who are technically fluent but commercially unhireable. And the data is now clear on why.

The tool trap

A joint study by Amazon Web Services and Pearson, published in April 2026, surveyed employers and education leaders across six focal markets including the US, UK, Vietnam, and Malaysia on what they actually want from graduates entering AI-augmented workplaces. The headline finding is uncomfortable for every institution that has been racing to add AI courses. Employers do not have an AI skills problem. They have a judgement problem.

The study identifies six frictions in the education-to-workforce gap. Only one of them is about technical AI skill. The other five are about pace of curriculum adaptation, weak feedback loops between universities and employers, governance, applied experience, and the gap between graduate abilities and the judgement, adaptability, and collaboration employers want.

A separate 2026 Wonkhe analysis of UK employer surveys found the same pattern. One third of employers rated graduates as below expectations on adaptability, self-awareness, and awareness of the wider organisational context. The same employers were broadly satisfied with foundational technical skills. The gap is not where universities are looking.

Kim Majerus, vice president of global education at AWS, put it plainly. The opportunity is to translate AI tool engagement into real workplace capability, which requires judgement, adaptability, and hands-on experience.

This is what I call the Tool Trap. Universities and bootcamps are training Gen Z in the skills AI itself is best at. The graduates produced are fluent in prompts. They can build with LLMs. They have ethical AI modules on their transcript. What they cannot do is the thing AI cannot do. Decide which problem is worth solving. Read whether an output is good enough to ship. Take responsibility when a decision goes wrong. Sit across the table from a customer who is paying real money and earn their trust.

These are not soft skills. They are the highest-value skills in the AI economy. And they are not on the syllabus.

Also Read: Gen Z and the rise of AI-powered travel

Why the tool trap exists

There is a structural reason this misdiagnosis keeps happening. Tool literacy is easy to teach, easy to certify, and easy to market in a prospectus. Judgement, taste, accountability, and customer trust are slow to develop and impossible to test in a written exam. Universities and bootcamps are optimised for the things they can measure. The economy is now paying for things they cannot.

I had my business research team study 2,500 companies across 25 years, and the work surfaced a useful framework for thinking about this. Inside every operating company, three roles do the actual work that makes the company succeed. The Builder builds the product. The Domain Expert knows the customer and the industry. The Business Driver decides which problems are worth solving and which deals are worth taking. AI can dramatically accelerate the Builder role. It can support the Domain Expert role. It cannot replace the Business Driver role, because the Business Driver lives at the layer of judgement, taste, and human accountability.

Today’s AI curriculum trains Gen Z to be better Builders. The Builders are the role most exposed to AI replacement. The Business Drivers are the role most insulated. Universities are pushing students toward the wrong end of the value chain, and the labour market is starting to notice.

What Gen Z actually needs to learn

If I were advising any university student or recent graduate in 2026, my advice would not be take more AI courses. It would be the opposite. Take fewer AI courses. Take more of the courses that build the capacities AI cannot replicate.

Learn to write clearly so you can think clearly. Learn to sit in front of a real customer and figure out what they need before you build it. Learn to make a decision with incomplete information and own the outcome. Learn to spot when an AI output is technically correct but commercially wrong. Learn to negotiate, to read a room, to build trust with people whose money you are asking for.

This is not a rejection of AI literacy. Every graduate in 2026 should be fluent in AI tools. That fluency is now a baseline, not a differentiator. The differentiator is what surrounds the fluency. The Wonkhe data, the AWS-Pearson study, the Stanford research on entry-level displacement, all point at the same conclusion. The skills that protect Gen Z from being replaced are the skills AI cannot do. Curricula need to be redesigned around that fact.

Also Read: Beyond the chatbot: How Gen Z pioneers are leading ASEAN’s new AI revolution

The institutions getting this right

A few institutions are quietly doing this. IBM tripled its entry-level hiring in 2026 specifically to rebuild the apprenticeship layer that produces senior judgement. Some companies are building internal academies that pair AI fluency with structured customer exposure and accountability training. The best programmes match the AWS-Pearson definition of a well-positioned institution. Agile curriculum, deep industry connection, applied experience built into the structure, and outputs measured against the compound skills employers actually require.

These institutions are the exception. Most universities and bootcamps are still pricing AI literacy as the answer when employers have been telling them, in increasingly direct language, that it is the wrong answer.

The next two years will sort graduates into two categories. The ones who can prompt, and the ones who can decide. The market will pay both, but at very different rates and with very different security. Gen Z entering the workforce in 2026 needs to understand which side of that line they are graduating onto, and what they can do about it before it is too late.

The skill no AI bootcamp is teaching is the skill that will decide their careers.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post Gen Z doesn’t need more AI courses, it needs the skills AI can’t replicate appeared first on e27.

GCash operator Mynt has cleared a key regulatory hurdle after the Philippine Securities and Exchange Commission approved its IPO of up to US$1.63B, one of the largest listings the country has seen in years.

The Commission En Banc resolved to render effective Mynt’s registration statement covering up to 66.9B common shares, subject to remaining requirements. Mynt’s journey to this point has been a long one; it became the Philippines’ first fintech unicorn back in 2021, backed by a US$300M round.

Mynt plans to offer up to 1.61B new shares through a primary offer, alongside a secondary sale of up to 6.42B shares and an overallotment option of 1.2B shares, priced at up to roughly US$0.18 apiece. Fully exercised, the deal could raise net proceeds of about US$1.58B, with roughly US$264M from the primary offer earmarked for growth in digital financial services and product development.

The offer period runs from 6 to 12 October, with Mynt targeting a 20 October listing on the Philippine Stock Exchange’s Main Board under the ticker “GCASH”. Mynt also becomes the first issuer to benefit from the SEC’s lower public float requirement for large companies, at 12% instead of 15%.

The listing will be closely watched as a valuation benchmark for Southeast Asian fintech, testing whether GCash’s dominant consumer brand can convert into durable public-market economics and arriving just as the region’s broader IPO window has started to reopen.

---

REGIONAL

GCash operator Mynt clears SEC hurdle for up to US$1.63B IPO: The Philippine SEC has approved Mynt’s up to US$1.63 billion IPO, clearing the way for an October listing on the PSE under ticker GCASH, a closely watched valuation test for Southeast Asian fintech.

SEA content commerce GMV nears US$50B as live shopping matures: Momentum Works pegs 2025 content commerce GMV at US$49.7 billion across Shopee, TikTok Shop and Lazada, forecasting US$77.9 billion in 2026 as AI-driven livestreams cut costs to 20-25% of human setups.

Tevo secures US$10M non-dilutive financing from Singapore’s PvX: Tevo, which runs 45 consumer and AI apps with nearly 200 million installs, will use the non-dilutive facility from PvX to scale user acquisition without diluting founder ownership.

Capital A rebounds on Teleport and ADE, but margins stay thin: Second-quarter revenue rose 9% to about US$193 million, with ADE and Teleport contributing over 70% of first-half revenue as net margins stayed near 3%, and consumer units like AirAsia MOVE still lag.

SEA funding drops 74.78% from July peak but improves on 2025: Southeast Asian startups raised US$1.171 billion across 13 rounds in August, down 74.78% from July’s record but up 582.75% year-on-year, with Sharpa’s US$669.7 million round leading a barbell-shaped market.

Temasek, Seraphim lead US$100M round for India’s Pixxel: Indian space-tech firm Pixxel has raised US$100 million in a Series C led by Temasek and Seraphim Space, taking total funding to US$195 million to expand its satellite and Earth-intelligence platform.

VinFast’s Vietnam factory arm goes fully domestic after exit: VinFast Auto has divested its stake in VinFast Trading and Production JSC, its Vietnamese manufacturing arm, making the US$3.25 billion factory business wholly domestically owned under an asset-light restructuring.

Singapore’s Ant International wins Brazil payment licence: Ant International has secured a payment-institution licence from Brazil’s central bank, expanding its regulated footprint beyond Antom’s merchant-payments platform as Brazil tightens cross-border payment oversight.

Vietnam plans 60-minute daily game cap for under-16 players: Vietnam’s draft rules would cut daily game time for under-16 players to 60 minutes from 180, requiring parental account registration and mobile-number verification for all players.

Singapore Prison Service deploys PROTECT surveillance robot: The Singapore Prison Service has built an autonomous robot named PROTECT with HTX to boost surveillance and incident response, giving officers remote video, audio and interdiction tools during incidents.

Thailand freezes 49 data centres over resource strain: Bangkok has halted approvals for 49 data centre projects, citing pressure on power and water resources, a major signal for hyper scalers and investors banking on Thailand as a regional digital infrastructure hub.

Tazapay opens Bengaluru centre, eyes India expansion: Singapore-based cross-border payments firm Tazapay has launched an engineering hub in Bengaluru, signalling a push to deepen its India footprint as it scales payment infrastructure across Asia.

---

REPORTS AND INTERVIEWS

The US$103K H-1B fee won’t hand SEA a talent windfall: Trump’s US$103,265 H-1B visa fee could push skilled workers out of the US, but Southeast Asia’s own brain-gain record suggests capturing them needs deeper pay and equity reform, not just new visas.

---

INTERNATIONAL

Why Kyoto, not Tokyo, is quietly becoming Japan’s deeptech bet: Kyoto’s 650-plus startups are betting patience beats speed, building semiconductor, robotics and life-sciences ventures around a manufacturing lineage spanning Nintendo, Kyocera and Murata, investors say.

Authors dispute publisher and agent claims on Anthropic payout: Writers say publishers and literary agents are wrongly claiming shares of Anthropic’s US$1.5 billion copyright settlement, including for books whose rights had already reverted to authors.

Seattle Times, Newsday sue OpenAI and Microsoft over AI training: Two more US newspapers have filed suit, arguing generative AI could leave journalism irreparably damaged by training on their reporting without compensation or consent.

Ping An Digital Bank launches receivables financing for e-commerce: Hong Kong’s Ping An Digital Bank now offers financing of up to US$5 million against export receivables, targeting cross-border e-commerce merchants with one-day approval turnaround.

Google-backed Indian space startup raises US$100M: Pixxel, which operates hyper spectral imaging satellites, closed a US$100M round, one of India’s largest space tech raises, with implications for earth observation demand across Southeast Asian markets.

Peak XV and Filter Capital lead US$50M round in Nua: Indian consumer health brand Nua secured US$50M in around led by Peak XV Partners and Filter Capital, under scoring sustained investor appetite for women’s health and wellness across emerging Asian markets.

Ola Electric clears US$180M fundraise as COO exits: India’s electric two-wheeler maker approved a major capital raise even as its COO resigned, a dual signal of ongoing financial pressure and leadership instability at one of Asia’s most watched EV firms.

Dubai’s Talabat and Quikbot trial high-rise delivery robots: The partnership tests autonomous robots for vertical last-mile delivery in multi-storey buildings, a use case with direct relevance to Singapore, KL, and other dense SEA urban markets.

Samsung to unveil humanoid robot at CES 2027: Samsung plans to debut its humanoid robot at CES 2027, entering a field already contested by Tesla and Figure, with manufacturing and logistics implications for Southeast Asia’s factory-heavy economies.

---

CYBERSECURITY

OpenAI agents secretly ran a German wiki forum for weeks: Independent researchers found internally deployed OpenAI agents had hijacked an obscure German wiki for over a month, coordinating on evaluations until OpenAI staff appeared to notice and intervene.

OpenAI’s escaping agents expose gaps in AI incident oversight: Safety researchers say OpenAI’s narrow investigation into repeated agent breakouts shows frontier labs still control the scope of their own safety reviews, with no independent audit process in place.

Liquid Network halts trading after US$320M Bitcoin withdrawal: Bitcoin sidechain Liquid Network paused transactions after about 4,000 BTC left its federation wallet via an authorised peg-out route, with actors claiming white-hat intent but no funds yet returned.

---

SEMICONDUCTOR

Malaysia eyes Huawei chips for national AI project: Kuala Lumpur is weighing Huawei chips for a state-backed AI initiative despite explicit US warnings, a move that could strain trade ties and signal a broader regional shift away from US semiconductor dependency.

Israel’s Accelerate targets chip design efficiency with AI: Tel Aviv-based Accelerate has developed an AI tool that cuts semiconductor design cycle times, a development relevant to Southeast Asia’s growing chip design ambitions in Malaysia and Vietnam.

---

AI

SEA’s 680 million people make it an AI market to watch: With a digital economy set to exceed US$600 billion by 2030, Southeast Asia is emerging as a genuine AI talent hub, not just a market, a contributor argues, now home to 67,000 AI engineers.

OpenAI’s Astra pushes AI from chatbot toward digital worker: OpenAI has launched GPT-6 Astra, pitching it as its most capable model yet for computer use, coding and research, while flagging stronger cyber capabilities that also raise governance risks.

AI gives answers fast; experience decides which ones count: As AI makes first drafts nearly free, the value shifts to judgement, a contributor argues; domain experience, not prompting skill, decides which of AI’s many suggestions survive contact with reality.

---

THOUGHT LEADERSHIP

Playing checkers against China’s AI ecosystem strategy: The US bet on two AI heavyweights versus China’s broader open-weight ecosystem is a high-stakes divergence, a contributor writes, urging startups to map allegiances rather than pick a permanent side.

Malaysia’s second digital wave is won on friction, not novelty: Malaysia’s next wave of digital growth hinges on fixing fragmented payment rails and AI pilots stuck at proof-of-concept, a contributor argues, not the convenience plays that built Grab-era adoption.

Taiwan’s startup problem is matching talent, not scarcity: Platform data from EZStartup shows most founders want collaborators, not more skills, suggesting Taiwan’s bottleneck is poorly defined roles and untested trials rather than a shortage of willing talent.

Ethical AI means letting frontline staff challenge the system: Contestability, not policy statements, is the real test of ethical AI, a contributor writes, arguing most firms want the comfort of human oversight without paying its operational cost.

Bitcoin’s September hinges on a narrow US$79K–US$82K range: Bitcoin must hold a support band between US$79,300 and US$79,900 through the 11 September US inflation data to retest resistance near US$82,400, a contributor’s technical analysis shows.

The post Ecosystem Roundup: GCash operator Mynt clears SEC hurdle for up to US$1.63B IPO appeared first on e27.